sensitivity property
How "sensitive" this Normalization
should be as an double
.
Clamped between 0..1
.
Influences the size of a buffer that stores past sensor samples.
That is to say, this sensitivity
is subtracted from 1.0
, then the
result is multiplied by 1000
and used as the maximum List
length
for a sensor samples buffer.
sensitivity
| calculation
| buffer length
1.0
=>(1.0 - 1.0) * 1000
=0
0.9
=>(1.0 - 0.9) * 1000
=100
(default)0.0
=>(1.0 - 0.0) * 1000
=1000
A larger sensitivity
, and thus smaller samples buffer means
new, strong outlier accelerometer data stands out more easily from the
average of the past samples in the buffer.
This will cause autocompensation on AcceleraxLayer
s (or XLayer
s)
to kick in more easily, delayed by a wait for delay Duration
.
A smaller sensitivity
also increases the size of an internal
List<List<double>>
buffer, so consider memory consumption.
Default is 0.9
for a buffer List
length of 100
.
With sensitivity == 1.0
, autocompensation would
consistently occur every delay Duration
.
Implementation
final double sensitivity;